ANATOMIC
PATHOLOGY
Original
D N A
A n a l y s i s
W i t h
Article
( P l o i d y )
I m a g e
o f
M o l a r
A n a l y s i s
T i s s u e
o n
P r e g n a n c i e s
P a r a f f i n
S e c t i o n s
IAN D. R. BARCLAY, BSC, LAITH DABBAGH, MSC, JEANNET BABIAK, MSC,
AND SIBRAND POPPEMA, MD, PHD, FRCP(C)
Hydatidiform moles can be subclassified based on their ploidy. In general, complete moles are diploid, and partial moles are triploid. The
standard method for the determination of DNA content is flow cytometric analysis. In this study, the authors investigated whether static cytometric analysis with the CAS 200 Image Analyzer (Cell Analysis Systems, Inc., Elmhurst, IL) with a software program designed for
quantitation of nuclear DNA content in tissue sections can be used to
classify moles. Tissue sections from 17 moles were analyzed with this
system, and the results were compared with those obtained with flow
cytometric analysis. It was found that cell selection was an important
factor. A high proportion of the hyperplastic trophoblast was in G2M.
Exclusion of these areas and measurement of the trophoblast lining the
villi only led to reliable results, and complete agreement between the
results of the two methods was obtained. The findings indicate that
cytometric analysis on tissue sections is a reliable alternative to flow
cytometric analysis for the designation of moles as diploid or triploid.
(Key words: Trophoblastic disease; DNA content: Hydatidiform moles)
Am J Clin Pathol 1993;100:451-455.
Hydatidiform moles can be defined as irregular products of
conception that result in placental maldevelopment. The placental villi are characteristically swollen, with abnormal trophoblast present. An embryo or fetus may be found.1 Moles have
been shown to exist in approximately 1 in 5,000 pregnancies2
and result from an unequal contribution of maternal and paternal chromosomes. ',3,4 They can be subgrouped into two categories consisting of partial and complete hydatidiform moles.
It is important to distinguish a complete mole from a partial
mole because partial moles are usually benign and without
danger of malignant transformation and metastasis; therefore,
they do not require chemotherapy.5"8 There have been rare
cases, however, in which partial moles developed into persistent trophoblastic tumors with myometrial invasion and persistent elevation of human chorionic gonadotropin levels, necessitating chemotherapy.910 In contrast, complete moles have
a higher risk of developing into a gestational trophoblastic tumor or choriocarcinoma, which require chemotherapy.2,5'8"
Because it is difficult to classify moles as complete or partial
solely on gross morphologic and microscopic features, in recent years pathologists have come to rely on quantitative DNA
analysis (ploidy) to aid in classification.5 Most complete moles
are diploid, whereas partial moles are triploid. The standard
method for determining DNA content isflowcytometric analysis. Recently, cytometric analysis systems, such as the CAS 200
Image Analyzer (Cell Analysis Systems, Inc., Elmhurst, IL),
have been developed. Image analysis combines computer software applications and microscopic image scanning techniques
to allow quantitative analysis of the DNA within individual
cells on cytologic specimens. The most recent improvement
consists of the development of a software program that allows
the quantitation of DNA in tissue sections that (by definition)
contain only slices of the nuclei. This allows quantitation of
DNA in specific cell types that can be defined by their morphologic characteristics and histologic location. Such a method
appears to be well suited for the quantitation of DNA in paraffin tissue sections of moles.
In this study, we compared the amount of DNA obtained by
image analysis on the CAS 200 with the results byflowcytometric analysis of 17 hydatidiform moles. The tissue samples used
in both methods were fixed in formaldehyde solution and embedded in paraffin.
MATERIALS AND METHODS
Paraffin-embedded tissue blocks of hydatidiform moles
from 17 patients were analyzed by flow cytometry and image
analysis. Both assays were run as blind independent studies,
and results were compared on completion of the study.
Slide Preparation
The specimens to be analyzed were cut into 5-/tm sections.
From the Department of Laboratory Medicine, Cross Cancer Before
Insti- relaxation in 4% buffered neutral formaldehyde for 30
tute. Edmonton, Alberta, Canada.
minutes, the slides were deparaffinized and rehydrated. The
slides then were stained with the CAS Quantitative DNA StainReceived June 19, 1992; revised manuscript accepted for publication
ing Kit (Cell Analysis Systems). After hydrolysis in 5 N HC1 for
January 12, 1993.
60 minutes, the slides were stained with Feulgen stain for 60
Address reprint requests to Dr. Poppema: Director of Laboratory
minutes. Reagents were rinsed for 30 seconds, 5 minutes, and
Medicine, Cross Cancer Institute, 11560 University Avenue, Edmon10 minutes; washing and decolorization with acid alcohol were
ton, Alberta, T6G 1Z2, Canada.
451
ANATOMIC PATHOLOGY
Original Article
452
was found that they were not of value in the analysis because
they always had a tetraploid DNA content. Cellular ploidy then
was determined by measurement of the nuclear optical density.
The DNA mass then could be determined with the formula m
= kOD, where k is a constant that incorporates the extinction
coefficient, wavelength of light, bandpass of light, and digitized
spot size, and OD is the optical density. The DNA index (DI) is
the ratio of the modal DNA content of the G0/G1 cells of the
sample divided by the modal DNA content of the diploid GO/
Gl reference cells. The image analysis system then generated a
histogram providing the DI of the primary peak.
The moles then were classified according to the DI of the
main peak. DNA indices ranging from 0.9 to 1.1 were considered diploid and DIs ranging from 1.4 to 1.6 as triploid. Tetraploid moles were not encountered in our study.
•sras*
FIG. 1. Hematoxylin and eosin-stained section of a mole showing villi
with a single layer of trophoblast cells and a solid area with hyperplastic
trophoblast (magnification, X40).
performed for five minutes. After decolorization, the slides
were dehydrated and cleared in xylene before they were coverslipped with a synthetic resin. Controls consisted of smears of
rat hepatocytes that represent a standard diploid and tetraploid
cell population.
Image Analysis
Image analysis was performed with the CAS 200 on 100-300
nonoverlapping, evenly distributed nuclei of trophoblast cells
surrounding the villi. The ALL DNA filter (CAS software; Cell
Analysis Systems) was used to select cells for analysis. Maternal
decidual cells were used as a diploid internal control. In all
cases, the thickness correction factor was applied. Regions with
hyperplastic trophoblast cells were measured separately, but it
Flow Cytometric Analysis
The molar specimen was dewaxed, rehydrated, and digested
enzymatically as described by Babiak and Poppema.12 After
resuspension in 0.5% (weight/volume) paraformaldehyde in
phosphate-buffered saline, the tissue was centrifuged. Samples
then were treated with 1 mL cold Triton X-100 (Sigma Chemical Company, St. Louis, MO) and incubated on ice to increase
the nuclear permeability. Residual RNA was removed, and the
DNA was stained by incubation with 50 Mg/mL propidium
iodide in phosphate-buffered saline for 30 minutes before analysis. The stained nuclear suspensions were analyzed on a FACScan flow cytometer (Becton Dickinson Immunocytometry
Systems, San Jose, CA) with a 488-nm argon laser for excitation. Before analysis, each sample was passed through a 25gauge needle with a syringe andfilteredthrough a 53-^m nylon
mesh filter. A minimum of 20,000 events were measured per
sample with low flow rates (12 jtL/minute) to maximize the
resolution. DNA indices and cell distributions then were calculated.
295
16
12
CO
o <n _
c
3
o
"5
u. <n c
3
o
o
=
8
a>
O
-
4_mi
i i i i I i i i i | i i i i | r i i i | i i i i |
200
400
600
800
1000
B
8
16
24
DNA Mass (Picograms)
32
FL2-A
FiG. 2. Comparison offlowcytometric analysis (A) and image analysis (B) of a triploid mole. There is a prominent G2M population in the image
analysis because all areas, including the hyperplastic trophoblast, were analyzed.
AJ.C.P. • October 1993
453
BARCLAY ET AL.
Cytometric DNA Analysis of Moles on Sections
produced more accurate results (Fig. 4). Of the 17 moles analyzed, 4 were found to be triploid and the remaining 13 were
diploid (Table 1). On review, five of the diploid cases were
diagnosed as complete moles; the other eight were believed to
show changes consistent with hydropic abortion. There was a
good correlation between the DIs obtained by image analysis
and flow cytometric analysis, resulting in 100% agreement in
the ploidy classification. Two of the five complete diploid
moles and none of the triploid cases had a persistent trophoblastic tumor develop that required chemotherapy during the follow-up period (range, 1-2 years).
The diploid classification by either method was relatively
straightforward because sharp peaks were visible in the 2N region of the histograms (Fig. 5). The 2N regions in some of the
CAS 200 histograms were relatively wide because of the smaller
sample size, but this did not affect the classification. The peaks
used to define a mole as triploid fell in the DI region of 1.5, but
the interpretation was not as apparent because there were variable diploid peaks and peak Mendings (Fig. 6).
c
O
O
"a)
o
8
16
24
DNA Mass (Picograms)
DISCUSSION
FIG. 3. Image analysis of hyperplastic trophoblast regions showing only
a prominent G2M population in these areas.
RESULTS
When all cells, including the thin cell layer of trophoblast
cells surrounding the villi and solid areas of hyperplastic trophoblast cells (Fig. I), were measured, the histograms showed a
considerable increase in the G2M phase that did not correlate
with the flow cytometric results (Fig. 2). When only the solid
areas were measured, we found an even spread through all cell
phase regions with a high G2M peak (Fig. 3). Analysis of only
the thin layer of trophoblast cells directly surrounding the villi
This study shows that image analysis and flow cytometric
analysis produce equivalent results with respect to quantitative
DNA analysis in hydatidiform moles. When image analysis
and flow cytometric analysis are compared, each has its own
advantages and disadvantages. Better resolution in the histograms is produced by flow cytometric analysis because it uses a
larger sample size than image analysis (20,000-30,000 cells as
compared with 200). Also, it uses whole nuclei instead of the
slices analyzed by image analysis; however, the sample is destroyed by the analysis, and sometimes cellular debris or a
small sample size will prevent an accurate measurement. Because the CAS 200 uses a tissue section (not a cell suspension)
as a sample, the sample is preservable. Image analysis allows
the operator to select cells individually, thus decreasing the
1
895
2
4
48"
36CO
o
{/)
-
c
_
3
0
O 24-
10 _
c
3
o
O
•5)
O
-
12'
4
rp-r
200
n
i' r i ' | 1 r r v i r i
800
400
i i i
1000
600
FL2-A
1 —
B
llwli j 1
r-1—1
8
16
24
DNA Mass (Picograms)
r
32
FIG. 4. Flow cytometric analysis (A) and image analysis (B) of a diploid mole. Theflowcytometric analysis shows a G2M peak that resulted from
including a large amount of hyperplastic trophoblast from this case. These hyperplastic areas were excluded in the image analysis.
Vol. 100-No. 4
454
ANATOMIC PATHOLOGY
Original Article
TABLE 1. COMPARISON OF FLOW CYTOMETRIC AND
IMAGE ANALYSIS DNA INDEX IN 17 CASES OF
HYDATIDIFORM MOLES
Patient
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Flow Cytometry
Index
2N
3N
2N
3N
2N
2N
2N
2N
2N
2N
3N
2N
3N
2N
2N
2N
2N
Image Analysis
Index
Agreement
1.02 (2N)
1.46 (3N)
1.10 (2N)
1.57 (3N)
1.10 (2N)
1.03 (2N)
1.04 (2N)
1.00 (2N)
1.09 (2N)
1.00 (2N)
1.56 (3N)
1.06 (2N)
1.57 (3N)
1.02 (2N)
1.09 (2N)
1.06 (2N)
1.09 (2N)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
number of cells needed and ensuring a specific cell population
in which cellular debris and irrelevant mixed cell populations
can be avoided. This enabled us to determine that the thin layer
of trophoblast cells surrounding the villi are the relevant cell
population; therefore, we could avoid measuring the hyperplastic G2M cells, which might lead to an inappropriate classification as tetraploid. In flow cytometric analyses, these cells are
included and show up in the G2M peak, which can be quite
prominent, as seen in figure 4A. Because of the large sample
size inflowcytometric analysis, this does not lead to problems
in interpretation and, in fact, can be used to estimate the proliferation fraction.
The dewaxing procedure and tissue thickness are important
variables in image analysis. The Feulgen stain is extremely sensitive to residual wax in the specimen, which results in blotchy
uninterpretable slides; thus, extra clearing time in xylene is
required. It is important to remember that the control slides do
not undergo the dewaxing process because these are smears,
not tissue sections, of rat hepatocytes. Thus, there is no control
in the dewaxing step. One might argue that only fresh or frozen
tissue should be used for image analysis because the fixation,
paraffin embedding, and deparaffinization are not standardized. Tissue thickness also must be monitored closely. Our sections were cut on a microtome set at 5 /um, but after the built-in
software of the CAS 200 was used to determine the tissue thickness (toggling), it was discovered that some sections were as
thick as 9 ^m. Our overall results were not affected because one
can make adjustments based on normal diploid cells. In addition, all of our sections were corrected for tissue thickness
through toggling, which did not change the actual results but
did produce sharper histograms; however, as the tissue thickness increased, it became more difficult to choose cells because
they stack on top of each other, limiting the number of cells
suitable for analysis.
The results of this study show that ploidy analysis of moles
on paraffin tissue sections is a good alternative toflowcytometric analysis. The ability to select the relevant cell population
makes up for the loss of resolution in analysis of only small
numbers of cells. It should be kept in mind that this does not
necessarily mean that DNA ploidy analysis on tissue sections
can replace flow cytometric analysis completely. The lack of a
precise resolution clearly will prevent applications in which
only small aberrations in the amount of DNA must be identified, as in most types of cancer. Analysis of whole cells in
smears or cytospin specimens or of whole nuclei that are isolated from paraffin blocks in a manner similar to that for flow
cytometric analysis may be a viable alternative and must be
studied in the future.
630
48
36c
3
o
O 24
"53
O
12-
I
I I I I | I I I I | I I I I | I I I I | I I I I |
/\ 0
200
400
600
800
1000
FL2-A
'
B
I
p.lll.11,1,.1.. •! .
8
16
24
DNA Mass (Picograms)
32
FIG. 5. Comparison of flow cytometric analysis (A) and image analysis (B) of a diploid mole. There are diploid peaks with only small G2M fractions
with both methods.
A.J.C.P. • October 1993
BARCLAY ET AL.
Cytometric DNA Analysis of Moles on Sections
455
75
CO
o
V)
"5
u.
(A
C
3
O
O
—
~
-
A
i i i i I i i i i I i r i i | i i i i | i i i i
A0
200
400
600
800
1000
B
FL2-A
8
16
24
DNA Mass (Picograms)
FIG. 6. Comparison offlowcytometric analysis (A) and image analysis (B) of a triploid mole. Prominent triploid peaks can be seen by both methods
in addition to the diploid peaks.
REFERENCES
1. Lage JM, Weinberg DS, Yavner DL, Bieber F. The biology of
tetraploid hydatidiform mole: Histopathology, cytogenetics and
flow cytometry. Hum Pathol 1989;20:419-425.
2. Fox H, ed. Obstetrical and Gynaecological Pathology. Third Edition. Edinburgh: Churchill Livingstone, 1991, pp 1045-1078.
3. Hall JG. Genomic imprinting: Review and relevance to human
disease. Am J Hum Genet 1990;46:857-873.
4. Surti U, Szulman AE, Wagner K, et al. Tetraploid partial hydatidiform moles: Two cases with a triple paternal contribution and a
92.XXXY karyotype. Hum Genet 1986;72:15-21.
5. Lage JM, Driscoll SG, Yavner DL, et al. Hydatidiform moles:
Application offlowcytometry in diagnosis. Am J Clin Pathol
1988;89:596-600.
6. Vassilakos P, Riotton G, Kajii T. Hydatidiform mole: Two entities. Am J Obstet Gynecol 1977; 127:167-170.
Benirschke K. Flow cytometry for all mole-like abortion specimens. Hum Pathol 1989;20:403-404.
Fisher RA, Lawler SD, Ormerod MG, et al. Flow cytometry used
to distinguish between complete and partial hydatidiform
moles. Placenta 1987;8:249-256.
Lage JM, Berkowitz RS, Rice LW, et al. Flow cytometric analysis
of DNA content in partial hydatidiform moles with persistent
gestational trophoblastic tumor. Obstet Gynecol 1991 ;77:111115.
10. Gaber LW, Redline RW, Mostouh-Zadeh M, et al. Invasive partial
mole. Am J Clin Pathol 1986;85:722-724.
Lawler SD, Fisher RA, Dent J. A prospective genetic study of complete and partial hydatidiform moles. Am J Obstet Gynecol
1991;164:1270-1277.
12. Babiak J, Poppema S. Automated procedure for dewaxing and
rehydrating of paraffin-embedded tissue sections for DNA flow
cytometric analysis of breast tumors. Am J Clin Pathol
1991;96:64-69.
Vol. 100 • No. 4